From 01f77ae25d80cdadba0e11e0455fc18708a2b63f Mon Sep 17 00:00:00 2001 From: Michael Yang Date: Thu, 4 Apr 2024 13:24:20 -0700 Subject: [PATCH] add command-r graph estimate --- llm/ggml.go | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/llm/ggml.go b/llm/ggml.go index 352c095f..071a36c3 100644 --- a/llm/ggml.go +++ b/llm/ggml.go @@ -327,7 +327,7 @@ func (llm GGML) GraphSize(context, batch int) (int64, bool) { } switch llm.KV().Architecture() { - case "gemma": + case "gemma", "command-r": return 4 * int64(batch) * int64(embeddingLength+uint64(vocabLength)), true case "phi2": return max(